Archive for January, 2018

GOAJ2: January update

Wednesday, January 31st, 2018

Here’s how things are going for GOAJ2: Gold Open Access Journals 2011-2016 and The Countries of OAWorld 2011-2016 and related stuff (all linked to from the project home page), through January 31, 2018 [noting that most of the last day of each month is missing because of how statistics are done]:

  • The dataset: 325 views, 58 downloads.
  • GOAJ2: 1,362 PDF ebooks (and two paperbacks), plus 655 copies of chapters 1-7 (C&I 17.4)
  • Countries: 479 PDF ebooks (no paperbacks)
  • Subject supplement (C&I 17.5): 1,175 copies

Looking at the rest of gold OA (beyond DOAJ), Gray OA 2012-2016 has now been downloaded 2,180 times, while Gray OA 2014-2017: A Partial Followup, released in October 2017, has 700 downloads.

 

GOAJ3: Second note

Wednesday, January 24th, 2018

I’m about 20% of the way through the first pass.

The only noteworthy observation at this point is a more emphatic version of one worrisome note from the earlier post:

  • Indonesian OA seems to have a malware problem. At this point, of 103 Indonesian OA titles checked, 32 are flagged by Malwarebytes Pro as either being malicious or having malware-infected segments. (Others: 9 of 98 in Romania, 3 of 180 in Brazil, two of nine in Ukraine, and one of 287 in Poland.) I’m hoping these will mostly resolve themselves on the second pass (or the third malware-only pass). I will not bypass Malwarebytes or Windows Defender protection. Period. Been there, done that, spent days fixing things.
  • Note that there were only 67 journals flagged as malware in last year’s pass. I now have 47 flagged out of the first 2,260–with more than 8400 left to go. I hope the rate goes down dramatically.

The requiem that isn’t

Sunday, January 21st, 2018

A week and a half ago, I started having problems with my computer while doing the marathon GOAJ3 scan of 10,800-odd OA journals. Once a day or so, usually late afternoon, first the external Sony display (where I had the “to do” spreadsheet) would go black, then the notebook display (usually showing Edge, although the “done” spreadsheet was also there but hidden, as was a Firefox session if I wanted a break from the scan) would go black, then one or both might come briefly back to life…and I’d eventually have to force a power-down and reboot. Sometimes I’d lose no work; sometimes 5-10 minutes (but after the second time, I started saving fairly often).

At first, I thought it might be the notebook itself, but after discussing it with the sensible person in the household–my wife–we concluded that it was most likely the external display, or possibly–unlikely–the cable.

Since replacing the cable (male VGA to male VGA) was really cheap, and since it was true that the old cable wasn’t a very secure fit on the notebook size (for some reason, Toshiba produced a notebook with only a VGA display connector–but didn’t seem to feel there was any reason to make it possible to secure the cable screws: there are no screwholes), I obtained a new cable. Still, we both assumed it was most likely the 12.5-year-old Sony SDM-HS95P display. After all, it was 12.5 years old (built in May 2005, and purchased shortly thereafter), and for the last 10.5 years it’s probably been on around 60-65 hours a week (that is, since OCLC decided I was of no further use: for the first two years, it probably got 30 hours a week or so. Add it up, and the display–expensive at the time (close to $1,000 as we remember it) and with spectacular quality for the time (look up the model number: it’s still one of the physically classiest displays around)–has probably been running for around 35,000 hours. I wouldn’t even assume a CFL backlight would last that long…

But…

I made one other change: instead of running the to-do spreadsheet full-screen, I ran it 22 rows tall (I normally do ten journals per pass, and move completed journal rows to the Done spreadsheet in 20-journal chunks), or about half-screen.

And when the blackout happened again, I realized that it was neither the cable nor the display: only the spreadsheet went black, while the Windows background picture stayed active. So, even after 12.5 years, the Sony continues to work (and to be a great display). I don’t see any rush to replace it whileit’s working: the 19″ (4×3 ratio) display fits the working space nicely and offers 1280×1024 resolution; to get the same vertical space and the same-size text on a newer (and certainly much cheaper!) 19×6 display, I’d probably have to go for at least a 24″ display, which would be awkwardly wide.

So the requiem or elegy for this long-surviving display will be put off for a while.

What’s the actual problem?

As far as I can tell. it’s either Excel 2013 or some interaction between Excel and Edge. (I’m using Edge for this pass because it seems faster than Chrome, has close-enough-to-equivalent autotranslate, and handles Ctrl-F, which I use a LOT, in a manner that makes it faster/easier for groups of journals with similar characteristics. I won’t make it my “standard” browser because it still won’t let me set a preferred typeface, so I’m dealing with the deluge of Arial/Helvetica/other dreary sans text on journal homepages.)

[Yes, I’ve done an Excel repair–but not a reinstall, because it’s not bad enough yet.]

Yesterday, another oddity arose, but this one turns out to be a blessing in disguise, sort of: Excel stopped converting cells containing URLs into hyperlinks by double-clicking. And here’s the thing: turns out that the Excel-activation-of-browser process, along with occasionally blanking out for a bit, was slower than what I’m doing now: copy the cell, open another Edge tab, paste the text, Enter. Four keystrokes and a touchpad move–just one keystroke more than the method I was using, and now I can inspect URLs without having them activate accidentally.

The blackout? Hasn’t happened again, yet, and perhaps the lack of Excel-to-Edge interaction has something to do with that. I still save early and often, and eventually I’ll try to figure out what’s going on and fix it.

But for now…back to the spreadsheets.

I’m leaving this one open for comments. Someone might be saying “I know what the problem is,” and if so you’re invited to comment. (Comments along the lines of “the problem is you’re still using Windows” will be cheerfully ignored.) And, yeah, maybe I’ll fire up LibreOffice and see how it does with the spreadsheet… [Added a bit later: Tried LibreOffice. At least with LibreOffice 5.4, there’s no such double-click-to-make-hyperlink functionality, so never mind.]


Update 2/8/18: It’s too early to say for sure, but I might have found the problem: Edge. Now running Chrome with Excel. On one hand, Find/ctrl-f across multiple tabs is much slower than with Edge; on the other, it’s nice not to have to look at all that bloody sans serif (since Chrome, like Firefox and unlike Edge, lets the user select a typeface. Most significantly (but it’s early yet), after the problem kept recurring every day or two, it now seems to be gone–and there are some suggestions online that Edge has or had stability problems. Sigh

Further update 2/13/18: Given the non-recurrence of stability problems, I’m now fairly sure that this is the problem–either some MS software doesn’t play nicely with other MS software or Edge just gets tired/cranky after lots of tab opening and closing and shows its edginess.

GOAJ3: First piece of the marathon

Saturday, January 13th, 2018

I’ve started in on GOAJ3, Gold Open Access Journals 2012-2017, which meansL:

  1. I’ve downloaded the DOAJ metadata (at 12:30 a.m. UMT on January 1, which was easy to do the afternoon of December 31).
  2. I’ve matched rows with the previous report where possible, since I don’t plan to recount 2012-2016 or revisit simplified subject assignments more often than necessary. (Matching is a multistage process.)
  3. I’ve tried to match remaining DOAJ rows with the “gray OA” spreadsheet to show cases where journals were either restored or have gained DOAJ status.
  4. I’ve started the actual data sweep, which will continue for somewhere between 10 more weeks and 88 more days (or more), after which comes revisiting cases needing revisits. [More plausible estimates, all assuming no significant personal or household crises: very optimistic: 10 weeks; mildly optimistic: 95 days; mildly pessimistic: 19 weeks.]
  5. In the process of starting tat, I’ve decided what’s worth trying to add to the analysis beyond the past two reports–which turns out to be mostly a crude measure of the transparency of APC status and amount. (As in, was the information reasonably readily available on the journal site or did I give up and rely on DOAJ?)

I’m a tenth of the way through–and that far only because of BioMed Central (I’m doing the data sweep alphabetically by publisher), with more than 200 journals that were exceedingly easy to get data for.

A couple of observations on that first tenth–actually a bit more:

  • Thanks to the group of hikers/walkers I go with on most Wednesday morning, I have eliminated the single “XT/couldn’t translate” omission from previous years. (The de facto leader of the Amblers, one of the three subgroups within the hiking group, is from India and I know the journal’s Indian. He agreed to take a look. It turns out to be in Hindi, not a language he reads–but a friend his *does* read Hindi and gave me the article coults.)
  • So far, it appears that Indonesia has a malware problem. 10 of 74 Indonesian journals checked so far are flagged by Malwarebytes as malware, The only other cases are five of 63 Romanian (I think one of those is actually an invalid security certificate) and two of 89 from Bosnia and Herzegovina. Naturally, those numbers will change–and I recheck all cases of malware (but I do not ever ignore Malwarebytes: problematic journals are one reason I use Malwarebytes Pro along with Windows Defender).

No real significance here; just a note along the (extended) way.

 

[Slowed down a bit because of a glitch in my two-display setup: after a $6 cable purchase, I’m now trying to determine whether it’s the cable or, more likely, a 12.5-year-old display. We shall see… ]